TLDR: On September 9, 2026, NinjaTrader launched its Innovation Lab, a team dedicated to building AI tools specifically for futures traders. The launch shipped three things at once: Nina, an AI companion that reviews a trader's own account activity and behavior; AI Strategy Builder, a conversational tool that turns plain-language rules into compiled, backtested NinjaScript strategies; and an MCP server, the first built for futures trading, that lets AI assistants such as Claude and ChatGPT read account data and place orders directly. Strategy Builder and MCP are both in beta, and NinjaTrader is explicit that traders remain responsible for whatever an AI agent does on their account.
Stop Executing Trades By Hand.
UMT Automator turns your TradingView or NinjaTrader strategy into automatic, hands-free execution — no code, no webhooks, no missed signals. Prefer a ready-made edge? Browse UMT's tested Strategies & Indicators built for both platforms.
Free 7-day trial on the Automator · No credit card required
The interesting part is not that NinjaTrader added a chatbot. It is that the launch spans three different jobs: coaching a trader's behavior, generating strategy code, and giving an AI agent live order-placement access. Those are three different risk profiles, and it is worth understanding what each one actually does before touching a live account.
What Is NinjaTrader's Innovation Lab?
The Innovation Lab is a dedicated team inside NinjaTrader focused on building AI natively into the futures trading experience, rather than bolting a generic chatbot onto the existing platform. NinjaTrader says the goal is to combine that AI expertise with a regulated, trader-controlled environment.
The launch was informed by a NinjaTrader survey of 1,500 active traders, which found that strategy development and skill-building were the top two areas traders wanted AI to help with. That survey result maps directly onto the two consumer-facing tools: AI Strategy Builder for strategy development, and Nina for skill-building and self-awareness.
Brian Weis, NinjaTrader's Chief Innovation & AI Officer, framed the bet this way: "The most important work in trading right now is where AI meets trading infrastructure. Nobody is doing that for futures traders." CEO Martin Franchi added that "retail investors want technology that matches the intensity and seriousness they bring to trading."
Nina: An AI Companion That Reviews Your Own Trading
Nina is built into the platform and focused on the trader, not the market. Instead of generating trade ideas or predicting price, Nina analyzes a trader's own account activity and behavior over time: which setups are working, which habits are costly, and where risk settings might need adjustment. NinjaTrader positions it as a feedback loop meant to build self-awareness, closer to a trading journal with a coach attached than a signal service.
That distinction matters for expectations. Nina is not a market-analysis tool and does not claim to forecast direction. It reviews what a trader already did and surfaces patterns in that history.
AI Strategy Builder: Plain-Language Rules to Compiled NinjaScript
AI Strategy Builder lets a trader describe entry, exit, risk, and filtering rules in plain English instead of writing NinjaScript by hand. NinjaTrader's help documentation describes the workflow in five steps:
- Describe the strategy in plain language, including entry and exit conditions, indicators, and risk parameters.
- The AI generates the corresponding NinjaScript code.
- NinjaTrader automatically compiles the strategy and runs a backtest in the same window.
- Review the results in the Strategy Analyzer panel.
- Refine the strategy through continued conversation without restarting the process.
Documented beta limitations are worth knowing before relying on it:
- English-only during the beta.
- A phased rollout, with a broader production release planned later.
- Indicators the tool does not yet support generate placeholder code that requires manual completion.
- Hourly and daily limits apply to how much a trader can generate.
- Any existing compile errors in the NinjaScript environment must be resolved before the tool will run.
NinjaTrader is direct that generated strategies are a starting point. Code review and independent validation come before backtesting conclusions are trusted, and backtesting conclusions come before any live capital.
The NinjaTrader MCP Server: Connecting Claude and ChatGPT to a Live Account
The most consequential piece of this launch is the MCP server. Model Context Protocol (MCP) is an open standard that lets an AI assistant connect to an external application's data and actions. NinjaTrader's implementation is, per the company, the first MCP server built specifically for futures trading, and it works over HTTPS with OAuth authentication, requiring no separate installation.
It supports any MCP-compatible AI client, which currently includes Claude Code, Claude Desktop, ChatGPT, Codex, and Cursor. Once connected, an agent can:
- Read account and portfolio data: net liquidation value, open P&L, positions, and trading history.
- Pull market research: live quotes, contract specifications, and margin requirements.
- Manage orders: place bracket orders, modify existing orders, and close positions.
- Configure risk and alerts: set daily loss limits, create price alerts, and adjust risk parameters.
NinjaTrader also ships 13 "trading skills" on top of the raw MCP tools, so a trader can phrase something like "size a trade on NQ with a 30-point stop" and let the skill select the right tools and do the math, rather than calling each tool individually.
What the MCP Server Actually Changes
This is a meaningfully different capability than a chatbot that answers questions about the market. A connected MCP client is not limited to describing what to do; depending on the permissions granted, it can place and manage real orders on a live account. NinjaTrader's own documentation is explicit about the resulting responsibility: "you are ultimately responsible for the trades your AI agent places." The documentation also warns that a connected agent "can make mistakes, misinterpret instructions, and act on incomplete or outdated information," and recommends testing in the Demo environment before connecting a Live account.
Before granting any AI client access to a funded account, review:
- Exactly which actions the client's approval settings allow it to take without confirmation.
- Whether order placement requires an explicit approval step per trade.
- Daily loss limits and other risk parameters, confirmed independently of the AI client's own summary.
- Demo-environment behavior across a range of prompts before any Live connection.
- How to immediately revoke MCP access if the agent behaves unexpectedly.
Nina vs. AI Strategy Builder vs. MCP: What Each One Does
| Tool | Primary job | Touches live orders? | Status |
|---|---|---|---|
| Nina | Reviews a trader's own account activity and behavior patterns | No | Built into the platform |
| AI Strategy Builder | Converts plain-language rules into compiled, backtested NinjaScript | No, by itself | Beta |
| MCP Server | Lets an external AI client read account data and place or manage orders | Yes, if permitted | Beta |
These tools solve different problems and carry different risk. Reviewing behavior, generating strategy code, and granting an AI agent order-placement access are not the same decision, even though they shipped under one announcement.
Where UMT Automator Fits
UMT Automator is built to automate eligible TradingView or NinjaTrader strategy activity through a supported, rules-based execution workflow. It is not the same product as NinjaTrader's MCP server: UMT executes a defined strategy's signals, rather than granting a general-purpose AI assistant open-ended access to place whatever orders it decides on.
If AI Strategy Builder helps a trader arrive at a rule set worth testing, the same discipline applies as with any other strategy source: backtest it, forward-test it on live data, paper trade the full execution path, and only then consider live automation. Review UMT's current requirements and check the free 7-day UMT Automator trial once a strategy has passed that process.
Frequently Asked Questions
What is NinjaTrader's Innovation Lab?
A dedicated team NinjaTrader launched on September 9, 2026 to build AI tools specifically for futures traders, led by Chief Innovation & AI Officer Brian Weis.
What does Nina do?
Nina reviews a trader's own account activity and trading behavior, highlighting effective and costly patterns and recommending risk adjustments. It does not analyze the broader market or generate trade signals.
Does AI Strategy Builder write working strategies automatically?
It generates compiled, backtested NinjaScript from a plain-language description, but NinjaTrader documents it as a starting point. Unsupported indicators produce placeholder code, and the tool is beta and English-only with usage limits.
Can Claude or ChatGPT place real trades through NinjaTrader?
Yes, if a trader connects that AI client to NinjaTrader's MCP server and grants order-placement permissions. NinjaTrader states the trader remains responsible for the agent's trades and recommends starting in the Demo environment.
Is the MCP server the same thing as UMT Automator?
No. The MCP server gives a general-purpose AI assistant direct account access. UMT Automator automates eligible, predefined strategy activity rather than granting open-ended agent access to an account.
Risk disclaimer: Futures trading involves substantial risk of loss and is not suitable for all investors. AI-generated strategies and AI agents connected to a live account can make mistakes, misinterpret instructions, or act on outdated information. Test any new tool in a demo or simulated environment before committing live capital, and verify NinjaTrader's current feature availability, beta status, and permission settings before use.
